Copilot for Manufacturing: 12-Wk Proof of Concept


Copilot for Manufacturing assist factory workers by leveraging OT and IT data to provide insights for automated root cause analysis, OEE analysis, document generation and industrial operations tasks

Manufacturing challenges solved

Manufacturing landscape has evolved with Industry 4.0 technologies, but factory workers still struggle with problem solving and effective data availability, having manufacturing data siloed in several separated OT and IT technologies, like MES - Manufacturing Execution System, IIOT – Industrial Internet of Things, WMS – Warehouse Management System, ERP - Enterprise Resource Management, CMMS - Computerized Maintenance Management Systems, S&OP - Sales & Operations Planning, DMS – Document Management System.

Cluster Reply, part of Reply Group, has created Copilot for Manufacturing, a Generative AI solution that analyzes vast amounts of structured and unstructured, real-time and historical data from various sources, thus supporting every industrial operation:

  • Maintenance: facilitates swift and precise identification of root causes, empowering proactive problem solving.
  • Operations: OEE problems are analyzed, casualized and proactively solved.
  • Document drafting: work and assembly instruction, compliance and QA documents, user and technical manual drafts are prepared by the generative capabilities.

Key benefits:

  • OEE increase: Enhances decision-making with real-time guidance, leading to improved OEE (Overall Equipment Effectiveness) and production efficiency, reduced downtime, cycle time and energy and material wastes.
  • Cost savings: Efficient issue resolution leads to significant cost reductions in labor and to savings related to increased availability and reduced production losses.
  • Continuous improvement: Helps identify and address inefficiencies and bottlenecks, fostering a culture of continuous improvement and productivity enhancement in line with LEAN principles.
  • Standardized work: Creates shared procedures and standardized work instructions ensuring process consistency, reduced variability and easier replication across workforce.
  • Empowering and skills development: Ease of knowledge access empowers frontline workers and enables skill development and knowledge sharing, also improving training and onboarding processes.

Cluster Reply Copilot for Manufacturing

By combining large language models (LLMs), Retrieval Augmented Generation (RAG) with industrial data lake and advanced analytics, Cluster Reply Copilot for Manufacturing is a comprehensive platform designed to transform manufacturing operations. It includes:

  • Comprehensive data layer leverages data lake (e.g. Microsoft Fabric, Databricks), direct APIs integrations (MES, CMMS, etc.) and Azure IoT Hub and Azure IoT Operations (to capture asset data, process it at the edge, and send it to the cloud) to create a unified data layer based on open standard and eventually aligned to ISA95, the leading industry standard for manufacturing operations management.
  • Copilot core combines Azure AI Search to facilitate search capabilities in OT and IT data sources and deliver more relevant results with Azure OpenAI to analyze Manufacturing data, documents, images and videos and generate insights and instructions. User feedback is collected for system improvement. Forecasting (production, energy, stock) and predictive maintenance insights are included by the means of traditional ML to guarantee explainability and reproducibility.
  • HMIs allow factory workers to interact through any touchpoint, from classic ones such as web and mobile apps delivered by Azure App Service and Microsoft Power Apps to chatbot and speech to text interaction in mobile devices, Microsoft Teams and Microsoft HoloLens.

What’s included

Proof of Concept timeline:

  • Use cases identification (1-2 weeks, optional): Workshops with business and IT stakeholders to identify the more valuable use cases and select priority one.
  • Use case discovery (1-2 weeks): analyze selected use case in detail, available data and expected benefits and metrics for success.
  • Implementation (6-8 weeks): Environment preparation and use case implementation.
  • Test and tuning (1-2 weeks): use case testing and fine tuning, benefits and metrics monitoring.
  • Use case(s) scaling (next weeks): roadmap for Generative AI scaling in the manufacturing processes.